The responsibility question has no clear answer yet:

• Developers create the models but cannot predict every use case or outcome
• Companies deploy the models in contexts the developers may not have intended
• Users rely on AI outputs they do not fully understand
• Regulators are years behind the technology and lack technical expertise to write effective rules

When a self-driving car causes an accident, who is liable? The car manufacturer, the software developer, the data providers, or the human who chose not to override? Current legal frameworks were not designed for autonomous decision-making systems.